Executive Summary
Industry classification is one of the most fundamental inputs in ESG and climate analysis. It determines how organisations: estimate emissions, assess transition risk, segment portfolios, and interpret exposure. Despite this, industry classification is often treated as a static administrative field rather than a dynamic analytical input. This creates significant challenges. When classification is inaccurate or outdated, it affects every downstream use case. This article explores why industry classification is critical, why traditional approaches fall short, and what a more effective model looks like.
1. Context: The Role of Industry in Climate Analysis
Climate risk is not evenly distributed across the economy. Different industries face different levels of: emissions intensity, regulatory pressure, transition exposure, and operational risk.
Industry classification provides the framework for understanding this variation. It allows organisations to: group entities, apply sector-level assumptions, and analyse exposure consistently.
2. The Problem with Traditional Classification
Most industry classification systems share common characteristics. They are: assigned manually, based on historical activity, rarely updated, and inconsistent across systems.
This creates several issues.
2.1 Static Assignment: Businesses evolve, but classifications often do not.
2.2 Lack of Granularity: Companies may operate across multiple activities.
2.3 Inconsistency Across Systems: Different systems may classify the same entity differently.
2.4 Lag in Reflecting Reality: Changes in business activity are not captured in real time.
3. Why This Matters
Industry classification underpins: emissions estimation, transition risk assessment, portfolio segmentation, and reporting outputs.
If classification is incorrect: emissions estimates become unreliable, risk assessments are distorted, and segmentation is flawed.
4. The Transition Risk Link
Transition risk is heavily dependent on industry. For example: energy-intensive sectors face greater pressure, while service-based sectors may face less direct impact.
Incorrect classification leads to: underestimation of risk and misallocation of resources.
5. The Scale Problem
At scale, classification errors multiply. Across thousands or millions of entities: inconsistencies become systemic, analysis becomes unreliable, and reporting becomes difficult to defend.
6. What Good Looks Like
Effective classification should be:
Activity-Based: Derived from actual business activity.
Dynamic: Updated as activity changes.
Consistent: Applied uniformly across datasets.
Scalable: Capable of covering large populations.
7. From Static Codes to Dynamic Insight
The shift required is from static classification to dynamic, data-driven categorisation.
This enables organisations to: reflect real-world activity, apply risk models accurately, and improve consistency.
Closing Insight
Industry classification is not simply a descriptive field - it is a critical analytical input. When it is inaccurate, every downstream use case is affected. Addressing this requires a structured approach that derives classification from real-world business activity and applies it consistently across large populations. This enables organisations to move from static categorisation to dynamic, scalable segmentation, supporting more accurate emissions estimation, risk assessment and reporting.